会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 2. 发明申请
    • SHADOW-PAGE DEFERRED-UPDATE RECOVERY TECHNIQUE INTEGRATING SHADOW PAGE AND DEFERRED UPDATE TECHNIQUES IN A STORAGE SYSTEM
    • SHADOW-PAGE DEFERRED-UPDATE RECOVERY TECHNIQUE INTEGRATING SHADOW PAGE AND DEFICH UPDATE TECHNIQUES IN A STORAGE SYSTEM
    • US20090164524A1
    • 2009-06-25
    • US12208014
    • 2008-09-10
    • Kyu Young WHANGYoung Koo LEEYi Reun KIMHyo Sang LIM
    • Kyu Young WHANGYoung Koo LEEYi Reun KIMHyo Sang LIM
    • G06F17/30
    • G06F11/1474G06F17/30345G06F2201/80G06F2201/82
    • Disclosed herein is a recovery method in a storage system that integrates shadow page and deferred update techniques. In our invention, for update operations that occur during a transaction, update of a page in the buffer is not immediately reflected to the database, but is deferred until the transaction is committed. Instead, we store the updated buffer page in the log volume in the unit of a page. When reading a page that does not exist in the buffer from the disk, we read the up-to-date page from the database or from the log volume. Here, the log table index is used to determine whether the up-to-date page exists in the database or in the log volume and to identify the position of the page stored in the log volume if the up-to-date page exists in the log volume. The proposed invention has the following excellent advantages. It is simple and efficient because the deferred update file does not have to be additionally accessed for correct query processing. Besides, there is no performance degradation because the position of an updated page in the database is not changed, and thus, the original clustering of logically related data is preserved.
    • 这里公开了一种将影子页和延迟更新技术相结合的存储系统中的恢复方法。 在本发明中,对于在事务期间发生的更新操作,缓冲区中的页面的更新不会立即反映到数据库中,而是延迟到事务提交。 相反,我们以页面为单位将更新的缓冲区页面存储在日志卷中。 当从磁盘读取缓冲区中不存在的页面时,我们从数据库或日志卷读取最新页面。 这里,日志表索引用于确定数据库或日志卷中是否存在最新页面,并且如果最新页面存在于日志卷中,则标识存储在日志卷中的页面的位置 日志卷。 所提出的发明具有以下优点。 它是简单和有效的,因为延迟更新文件不需要额外访问以进行正确的查询处理。 此外,由于数据库中更新的页面的位置没有改变,所以不会降低性能,因此保留了逻辑相关数据的原始聚类。